July 04, 2025
AI Agents Are Winning Hearts, and Here’s the Proof

Over half of users (63%) now believe AI agents can resolve their issues accurately, which is a dramatic increase from just 36% before seeing one in action. A new study from Intercom reveals how consumer attitudes toward AI-powered customer service can shift dramatically after seeing a modern AI Agent in action. The research, based on responses from over 1,000 end-users in the U.S., focused on sentiment, trust, accuracy, and overall optimism around AI in support interactions.
Sentiment Shift After Demo Exposure
Before seeing the AI Agent in action, only 40% of respondents said they had a positive perception of AI in customer service. Around 25% expressed negative sentiment. However, after watching a short demo that showcased the AI Agent handling real support tasks, positive sentiment rose by 20 percentage points while negative sentiment fell by 11 points, down to just 14%.
This shift shows that user scepticism about AI is usually tied to outdated experiences with basic chatbots. Demonstrating how AI can now handle full conversations and resolve issues autonomously has a measurable impact on perception.
Increased Trust in AI Capabilities
Trust levels saw a similar improvement. Initially, only 39% of respondents said they trusted AI to effectively solve their customer service problems. After the demo, that number rose to 57%. Additionally, distrust levels dropped from 35% to 18%.
Intercom’s report highlights that transparency and control are critical to building on this trust. The company recommends implementing safeguards such as defined content sources, custom responses for sensitive topics, and visible source citations to ensure the AI remains reliable and on-brand.
Another area the study explored was how users perceive AI’s ability to provide accurate answers. Based on previous experiences, just 36% believed AI would make it easier to get correct information. After viewing the demo, that number increased to 63%, marking a 27-point improvement.
To maintain this level of perceived accuracy, the report advises companies to build robust knowledge foundations and leverage AI Agents to triage incoming issues and collect context before escalating to human agents if needed.
Faster Resolution Times
In addition, the speed of support resolution was also examined. Before the demo, 38% believed AI would help resolve issues more quickly. That increased to 63% after seeing the AI Agent in action — a 25-point improvement. Users who had previously assumed AI might slow down interactions appeared more confident after observing automated workflows and real-time problem-solving.
The report outlines several examples where AI Agents can independently manage multi-step resolution processes. These include processing returns, updating account details, and verifying identity, tasks that often don’t require human intervention when the AI is well-trained and connected to backend systems.
When asked whether AI would improve customer service in the future, only 42% of users agreed before the demo. That rose to 56% afterwards. Meanwhile, the share of users who believed AI would make things worse dropped from 31% to 13%.